Predicting Happiness State Based on Emotion Representative Mining in Online Social Networks

Online social networks (OSNs) have become a major platform for people to obtain information and to interact with their friends. People tend to post their thoughts and activities online and share their emotions with friends, which provides a good opportunity to study the role of online social networks in happiness spreading and mutual influence among the users. In this paper, we propose a framework to study the influence of happiness in OSNs. We first quantify the happiness states of users by analyzing their daily posting texts, and then conduct the statistical analysis to show that users’ happiness states are influenced by their social network neighbors. Since the influence of each individual is unequal, we develop a regression model and a greedy algorithm to detect the high influence users known as emotion representatives. By using a small number of detected emotion representatives as features to train prediction models, we show that it achieves good performance in predicting the happiness states of the whole online social network users.